---
title: PixelRefer-7B
canonical_url: "https://www.modelscope.cn/models/DAMO_Academy/PixelRefer-7B"
md_url: "https://www.modelscope.cn/models/DAMO_Academy/PixelRefer-7B.md"
repository: DAMO_Academy/PixelRefer-7B
last_updated: 2025-11-26
license: apache-2.0
model_type:
  - pixelrefer_qwen2
architectures:
  - PixelreferQwen2ForCausalLM
base_model:
  - DAMO-NLP-SG/VideoLLaMA3-2B-Image
base_model_relation: finetune
library_name:
  - pytorch
frameworks:
  - pytorch
language:
  - en
downloads: 101
stars: 0
tags:
  - "multimodal large language model"
  - "large video-language model"
---

# PixelRefer-7B

> PixelRefer-7B - DAMO_Academy 在 ModelScope 开源的模型。PixelRefer: A Unified Framework for Spatio-Temporal Object Referring with Arbitrary Granularity

DAMO_Academy/PixelRefer-7B 是 ModelScope 魔搭社区上的机器学习模型，采用 apache-2.0 许可，基于 DAMO-NLP-SG/VideoLLaMA3-2B-Image 构建。

- **Repository**: DAMO_Academy/PixelRefer-7B
- **License**: apache-2.0
- **Base model**: DAMO-NLP-SG/VideoLLaMA3-2B-Image
- **Tags**: multimodal large language model, large video-language model
- **Downloads**: 101
- **Stars**: 0
- **Last updated**: 2025-11-26

Source: https://www.modelscope.cn/models/DAMO_Academy/PixelRefer-7B

---

<p align="center">
    <img src="https://cdn-uploads.huggingface.co/production/uploads/64a3fe3dde901eb01df12398/CiwESNwyy7VTooOifRgiQ.png" width="70%" style="margin-bottom: 0.2;"/>
<p>


<h3 align="center"><a href="http://arxiv.org/abs/2510.23603" style="color:#4D2B24">
PixelRefer: A Unified Framework for Spatio-Temporal Object Referring with Arbitrary Granularity</a></h3>

<div align=center>

![Static Badge](https://img.shields.io/badge/PixelRefer-v1-F7C97E) 
[![arXiv preprint](https://img.shields.io/badge/arxiv-2510.23603-ECA8A7?logo=arxiv)](https://arxiv.org/abs/2510.23603) 
[![Dataset](https://img.shields.io/badge/Dataset-Hugging_Face-E59FB6)](https://huggingface.co/datasets/DAMO-NLP-SG/VideoRefer-700K) 
[![Model](https://img.shields.io/badge/Model-Hugging_Face-CFAFD4)](https://huggingface.co/collections/Alibaba-DAMO-Academy/pixelrefer) 
[![Benchmark](https://img.shields.io/badge/Benchmark-Hugging_Face-96D03A)](https://huggingface.co/datasets/DAMO-NLP-SG/VideoRefer-Bench) 


[![Homepage](https://img.shields.io/badge/Homepage-visit-9DC3E6)](https://circleradon.github.io/PixelRefer/) 
[![Huggingface](https://img.shields.io/badge/Demo-HuggingFace-E6A151)](https://huggingface.co/spaces/lixin4ever/PixelRefer) 
</div>

            
## 📰 News
* **[2025.10.28]** 🔥We release PixelRefer.
* **[2025.6.19]** 🔥We release the [demo](https://huggingface.co/spaces/lixin4ever/VideoRefer-VideoLLaMA3) of VideoRefer-VideoLLaMA3, hosted on HuggingFace. Feel free to try it!
* **[2025.6.18]** 🔥We release a new version of VideoRefer([VideoRefer-VideoLLaMA3-7B](https://huggingface.co/DAMO-NLP-SG/VideoRefer-VideoLLaMA3-7B) and [VideoRefer-VideoLLaMA3-2B](https://huggingface.co/DAMO-NLP-SG/VideoRefer-VideoLLaMA3-2B)), which are trained based on [VideoLLaMA3](https://github.com/DAMO-NLP-SG/VideoLLaMA3).
* **[2025.4.22]** 🔥Our VideoRefer-Bench has been adopted in [Describe Anything Model](https://arxiv.org/pdf/2504.16072) (NVIDIA & UC Berkeley).
* **[2025.2.27]** 🔥VideoRefer Suite has been accepted to CVPR2025!
* **[2025.2.18]**  🔥We release the [VideoRefer-700K dataset](https://huggingface.co/datasets/DAMO-NLP-SG/VideoRefer-700K) on HuggingFace.
* **[2025.1.1]**  🔥We release [VideoRefer-7B](https://huggingface.co/DAMO-NLP-SG/VideoRefer-7B), the code of VideoRefer and the [VideoRefer-Bench](https://huggingface.co/datasets/DAMO-NLP-SG/VideoRefer-Bench).




## 🌏 Model Zoo
| Model Name      | Visual Encoder | Language Decoder | 
|:----------------|:----------------|:------------------|
| [PixelRefer-7B](https://huggingface.co/Alibaba-DAMO-Academy/PixelRefer-7B) | [VL3-SigLIP-NaViT](https://huggingface.co/DAMO-NLP-SG/VL3-SigLIP-NaViT) | [Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) |
| [PixelRefer-2B](https://huggingface.co/Alibaba-DAMO-Academy/PixelRefer-2B) | [VL3-SigLIP-NaViT](https://huggingface.co/DAMO-NLP-SG/VL3-SigLIP-NaViT) | [Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) |
| [PixelRefer-Lite-7B](https://huggingface.co/Alibaba-DAMO-Academy/PixelRefer-Lite-7B) | [VL3-SigLIP-NaViT](https://huggingface.co/DAMO-NLP-SG/VL3-SigLIP-NaViT) | [Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct) |
| [PixelRefer-Lite-2B](https://huggingface.co/Alibaba-DAMO-Academy/PixelRefer-Lite-2B) | [VL3-SigLIP-NaViT](https://huggingface.co/DAMO-NLP-SG/VL3-SigLIP-NaViT) | [Qwen2.5-1.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-1.5B-Instruct) |

## 📑 Citation

If you find VideoRefer Suite useful for your research and applications, please cite using this BibTeX:
```bibtex
@article{yuan2025pixelrefer,
  title     = {PixelRefer: A Unified Framework for Spatio-Temporal Object Referring with Arbitrary Granularity},
  author    = {Yuqian Yuan and Wenqiao Zhang and Xin Li and Shihao Wang and Kehan Li and Wentong Li and Jun Xiao and Lei Zhang and Beng Chin Ooi},
  year      = {2025},
  journal   = {arXiv},
}

@inproceedings{yuan2025videorefer,
  title     = {Videorefer Suite: Advancing Spatial-Temporal Object Understanding with Video LLM},
  author    = {Yuqian Yuan and Hang Zhang and Wentong Li and Zesen Cheng and Boqiang Zhang and Long Li and Xin Li and Deli Zhao and Wenqiao Zhang and Yueting Zhuang and others},
  booktitle = {Proceedings of the Computer Vision and Pattern Recognition Conference},
  pages     = {18970--18980},
  year      = {2025},
}

```